The relationship between illicit drug use and pregnancy outcome has been controversial. Self-reported drug use is notoriously inaccurate, and selective screening of blood and/or urine often results in biased ascertainment of drug use, as the reason for screening may be as important a risk factor as the drug use itself. There have been a few systematic large-scale studies of drug use at specified times during pregnancy in a defined cohort of women, and even fewer that have collected detailed data on other risk factors for adverse pregnancy outcome. In this project, serum obtained from approximately 7,500 women on entry to the Vaginal Infections and Prematurity Study (23-26 weeks) were screened for marijuana and cocaine metabolites. In addition, serum obtained at delivery of low birth weight, preterm and a random sample of term infants was screened, as was serum obtained at 31-36 weeks from a random sample of uncomplicated pregnancies and serum obtained from women experiencing abruptio placenta and placenta previa. The Vaginal Infections and Pregnancy Study collected extensive data on a variety of confounding variables, and enabled the most thorough analysis yet of the effect of cocaine and marijuana use on pregnancy outcome.